AU2010251701A1 - Water analysis - Google Patents

Abstract

A method of determining chemical oxygen demand in saline water samples containing chloride ions by a photoelectrochemical method in which the photo electrode is a titanium dioxide sensor treated by deposition of a noble metal oxide preferably silver(l) oxide or palladium, (II) oxide.

Description

WO 2010/132957 PCT/AU2010/000621 1 WATER ANALYSIS Field of the Invention This invention relates to a method for determining oxygen demand of water using photoelectrochemical cells. In particular, the invention relates to a 5 photoelectrochemical method of determining chemical oxygen demand of water samples having high chloride content such as sea water. Background to the Invention Nearly all domestic and industrial wastewater effluents contain organic 10 compounds, which can cause detrimental oxygen depletion (or demand) in waterways into which the effluents are released. This demand is due largely to the oxidative biodegradation of organic compounds by naturally occurring microorganisms, which utilize the organic material as a food source. In this process, organic carbon is oxidised to carbon dioxide, while oxygen is consumed 15 and reduced to water. Standard analytical methodologies for the determination of aggregate properties such as oxygen demand in water are biochemical oxygen demand (BOD) and chemical oxygen demand (COD). BOD involves the use of heterotrophic microorganisms to oxidise organic material and thus estimate 20 oxygen demand. COD uses strong chemical oxidising agents, such as dichromate or permanganate, to oxidise organic material. BOD analysis is carried out over five days and oxygen demand determined by titration or with an oxygen probe. COD measures dichromate or permanganate depletion by titration or spectrophotometry. 25 Application W02004/088305 discloses a photoelectrochemical method of detecting chemical oxygen demand as a measure of water quality using a titanium dioxide nanoparticulate semiconductor electrode. Titanium(IV) oxide (TiO 2 ) has been extensively used in photooxidation of organic compounds. TiO 2 is non photocorrosive, non-toxic, inexpensive, relatively easily synthesised in its highly 30 active catalytic nanoparticulate form, and is highly efficient in photooxidative degradation of organic compounds. This method is satisfactory for the analysis of water and wastewater samples which contain very low levels of the chloride ion C (say, [Cl] < -20 mg/L). When the chloride ion concentration in the sample to WO 2010/132957 PCT/AU2010/000621 2 be analysed exceeds this level the COD measurement may suffer increased uncertainty due to interference of the chloride ion in the electrochemical measurement process. A problem encountered in conducting assays using this method is dealing 5 with interference from competing oxidisable chemical species other than organic carbon. Filtration of samples reduces interference from many species but the presence of chloride still remains a significant interference that must be dealt with. The conventional dichromate COD detection method deals with chloride interference by chemically removing the chloride ions. The principle is to add a 10 chemical that can form compounds with C that are not oxidised by the dichromate ion, as exemplified by the following reactions Hg 2 + (aq) + 2 C(atq -- HgC 2 (aq)s Ag*(ad + CF(aq) - AgCI4(so,d), 15 Addition of Ag* as the ionic salt Ag 2

SO

4 , removes C ion from solution by precipitating out as AgCl(s). This is not oxidised to a large extent by the dichromate ion (Cr 2 0 7 2 -). Addition of Hg 2 + as the ionic salt HgSO 4 , to a solution containing Cr, results in the formation of mercuric chloride (HgC 2 ). HgCl 2 is not an ionic salt but rather is a triatomic molecule. When dissolved in water, the Cl atoms remain 20 complexed to the Hg atom. The HgC 2 molecule in solution is resistant to oxidation by the dichromate ion (Cr 2 0 7 2 -). Because of its high degree of solubility in water HgCl 2 is one of the most toxic forms of mercury known, and therefore presents a considerable toxic waste disposal problem. The conventional dichromate COD method involves the use of expensive and toxic 25 chemicals requiring careful disposal. For online applications, the system will need a sophisticated component to achieve in situ separation of precipitated AgCI, which, on one hand will significantly undermine the accuracy and reliability of the system, and on the other hand will increase both the capital and operational costs. For both Industry and environmental management, there is considerable need to 30 be able to measure the levels of organic content (or contamination) in water containing low to high levels of chloride (e.g. in sea water). Chloride is a problem for organic content measurement in aqueous samples as current methods of analysis can't easily distinguish between organic and chloride content, without WO 2010/132957 PCT/AU2010/000621 3 resorting to the use of toxic mercury. Historically, the US EPA has approved a particular method of dichromate digestion for COD measurement. The EPA approved method requires the addition of mercury and silver salts to mask the chloride interference prior to analysis. Often 5 this is accompanied by sample dilution in order to fully prevent unwanted oxidation of the chloride ions and to provide a solution capable of being analysed spectrophotometrically or by potentiometric titration. Thus low level organic content is not easily measured with any accuracy. This method also produces toxic salts that require specific materials handling procedures for disposal. 10 Another established method for quantifying organic contamination of water is total organic carbon (TOC) measurement. Difficulties in the presence of chloride also exist for TOC in that for persulphate based oxidation, false positives arise due to the inability to distinguish between chloride ions and the organic compounds in the solution. For high temperature oxidation based TOC, the chloride salts calcine and 15 can cause instrumental errors through degraded operating performance, thus leading to high maintenance requirements for the technique and an inability to conduct regular on-line measurements. W02007/016740 discloses an improvement in the photoelectrochemical method of detecting chemical oxygen described in W02004/088305 which deals with the 20 interference by chlorine. In water samples containing chloride ions above 0.5mM concentration in which the samples are diluted and a known quantity of an organic substance is added to the diluted sample which is then subjected to an assay by a photoelectrochemical method using a titanium dioxide photoactive nanoparticulate semiconductor electrode and the chemical oxygen demand is measured in the 25 same manner as disclosed in W02004/088305, except that a known concentration organic solution is used to obtain the blank for calculation of net charge Qnet . With this organic addition method, the analytical signal is generated in exactly the same way as the photoelectrochemical method disclosed in W02004/088305. However, this addition method requires some sort of prior knowledge of the range 30 of chloride concentration and organic contents in the samples in order to determine how much organic material to add to ensure total suppression of the chloride interference, and thus an accurate COD measurement. In many cases, this is not a serious issue, however in environments such as sea water where the chloride WO 2010/132957 PCT/AU2010/000621 4 content is high, high levels of organic addition and sample dilution are then required and this has an impact on the sensitivity achievable. Patent application W02009/049366 discloses an enhancement of the method disclosed in W02004/088305 which successfully overcomes the problem 5 presented by chloride ion interference. That enhancement was to irradiate the TiO 2 sensor surface with pulsed-UV radiation rather than with continuous-UV radiation, as is done in the method of W02004/088305. The pulsed-UV method facilitates the measurement of COD in water and wastewater samples with chloride concentrations ranging up to at least that found in natural seawater, (i.e., 10 [Cl~] = 21,000 mg/L). One disadvantage of the pulsed-UV method is that the time required for an analysis in this mode is considerably longer (say, by a factor of approximately three) than is required for the normal continuous-UV method used for chloride-free samples. There is a significant body of literature in both the scientific and patent domains, 15 concerning the deposition onto TiO 2 nanoparticles of noble metal nanoparticles in their zero (i.e., metallic) oxidation state (e.g., Ag*, Pt*, Pd 0 and Au*) or in higher oxidation states (e.g. Ag+ 1 , Pt+ 2 and Pt* 4 , Pd+ 2 and Au+'). Such treatment of TiO 2 has frequently been observed to increase the efficiency of (both the rate of and the completeness of) its oxidation of organics in bulk water or wastewater, relative to 20 the efficiency observed when using non-treated TiO 2 . USA patent 5872072 discloses a catalytic composition useful for decomposing malodorous compounds which includes titanium dioxide and an antimicrobial metal selected from silver copper and zinc. However, the literature is silent on modifications to TiO 2 which modify its photoelectrochemical behaviour with respect to the chloride ion. 25 It is an object of this invention to provide a simpler means of dealing with chloride interference than those currently known. Brief Description of the Invention To this end the present invention provides a method of determining chemical 30 oxygen demand in water samples containing chloride ions by a photoelectrochemical method in which the photo electrode is a titanium dioxide sensor treated with a noble metal compound. The noble metal is selected from the group of gold, palladium, platinum and WO 2010/132957 PCT/AU2010/000621 5 preferably silver. After treatment the titanium dioxide sensor surface includes an oxide of the noble metal and this is preferably a silver or palladium oxide. The Ag 2 0:TiO 2 and the PdO:TiO 2 density ratio of the resulting A920-TiO 2 or PdO TiO 2 composite material is preferably controlled by manipulating the deposition 5 parameters within the range of 0.01 to 0.4 preferably 0.05 to 0.15. This invention is partly predicated on the discovery that interference of chloride ions in the method disclosed in W02004/088305 (referred to as PeCOD* analysis) for measuring COD in water and wastewater manifests in at least two distinct ways, which are (a) signal suppression; and (b) signal tailing. These terms both 10 refer to ways in which the oxidation profile measured electrochemically at the nanoparticulate-TiO2 working electrode, (i.e., the /wo vs. time signal recorded by the instrument) is distorted, so as to yield an error in the calculation of COD from this profile. This invention is also predicated on the discovery that a TiO 2 sensor which has 15 been treated by the deposition of silver(l) oxide (Ag 2 0), will be far less sensitive to the presence of chloride ions in the water sample than a TiO 2 sensor which has not undergone such treatment with Ag 2 0. Using such an Ag 2 0-deposited TiO 2 sensor facilitates the rapid analysis of chloride containing samples using continuous (rather than pulsed) UV irradiation. This advantage applies to samples 20 in the PeCOD* measurement cell with chloride ion concentrations ranging up to at least [C]=700 mg/L. This yields a significant saving in time for the analysis of chloride containing samples compared to employing the pulsed-UV method of PCT2008/001529 mentioned above. In addition to this advantageous property, Ag 2 0-treated TiO 2 COD sensors are 25 found to linearise the instrument response function (i.e., measured Qnet vs. sample COD) of the PeCOD* analyser which, with non-Ag 2 0-treated TiO 2 is significantly (and reproducibly so) nonlinear. A linear instrument function facilitates the employment of simpler calibration protocols without risking the introduction of systematic measurement errors due to nonlinearity. 30 A further advantage is that Ag 2 0-treated TiO 2 COD sensors are found to render the PeCOD* analysis method immune to interference from dissolved carbon dioxide in the water sample. In the PeCOD* method dissolved carbon dioxide (C02) in a sample of water or wastewater can yield a spuriously high COD reading WO 2010/132957 PCT/AU2010/000621 6 when using a non-Ag 2 0-treated TiO 2 COD sensor. Although not a very large effect, it can be significant particularly if a small water sample has been left open to the atmosphere for a long time and has been allowed to come to a concentration equilibrium with atmospheric CO 2 . 5 A final advantage is that combining the Ag20-treatment of the TiO 2 sensor with the pulsed-UV approach yields a new pulsed method for analysis of samples containing chloride ranging up to at least [C[~]=21,000 mg/L, which is significantly faster than that disclosed earlier. Upon absorption of light by the TiO 2 photocatalyst, electrons in the valence band 10 are promoted to the conduction band (ecb~) and "photoholes" are left in the valence band (hvb*). The photohole is a very powerful oxidizing agent (+3.1 V) that will readily lead to the seizure of an electron from a species adsorbed to the solid semi-conductor TiO 2 . Thermodynamically, both organic compounds and water can be oxidized by the photoholes or surface trapped photoholes but usually organic 15 compounds are more favourably oxidized, which leads to the mineralization of a wide range of organic compounds. This is described in application W02004/088305 the contents of which are incorporated herein by reference. Owing to the strong oxidation power of photoholes, photocatalytic oxidation of organic compounds at the TiO 2 electrode leads to stoichiometric oxidation 20 (degradation) of organic compounds as follows: CHm 0,Xk Nj SiPh + b 0 2 -- nCO 2 + m-k-3j-2i-3h H 2 0 + kHX + jNH 3 + iH 2

SO

4 + hH 3

PO

4 where b = n + m-k-3j-2i-3h - + 2i + 2h 4 2 and where X represents a halogen atom. The numbers of carbon, hydrogen, 25 oxygen, halogen, nitrogen, sulphur, and phosphorous atoms in the organic compound are represented by n, m, e, k, j, i and h, respectively. In order to minimize the degradation time and maximize the degradation efficiency, the photoelectrochemical catalytic degradation of organic matter is preferably carried out in a thin layer photoelectrochemical cell. This process is analogous to 30 bulk electrolysis in which all of the analytes are electrolysed and Faraday's Law can be used to quantify the concentration by measuring the charge passed if the charge/current produced is originated from photoelectrochemical degradation of WO 2010/132957 PCT/AU2010/000621 7 organic matter. That is: Q = Jidt =nFVC where n refers to the number of electrons transferred during the photoelectrocatalytic degradation, and i is the photocurrent from the oxidation of 5 organic compounds. F is the Faraday constant, while V and C are the sample volume and the concentration of organic compound respectively. The measured charge, Q, is a direct measure of the total amount of electrons transferred that result from the complete degradation of all compounds in the sample. Since one oxygen molecule is equivalent to 4 electrons transferred, the 10 measured Q value can be easily converted into an equivalent 02 concentration (or oxygen demand). The equivalent COD value can therefore be represented as: COD (mg / L of 0 2 ) = -x 32000 4FV This COD equation can be used to quantify the COD value of a sample since the charge, Q, can be obtained experimentally and for a given photoelectrochemical 15 cell, the volume, V, is a known constant. In another aspect the present invention provides a photoelectrochemical assay apparatus for determining oxygen demand of a water sample which consists of a) a flow through measuring cell; b) a photoactive titanium dioxide working electrode which has been treated 20 by deposition with a noble metal oxide, preferably silver(l) oxide or palladium(ll) oxide and a counter electrode disposed in said cell; c) a UV light source, adapted to illuminate the photoactive working electrode either continuously or in pulses; d) control means to control the illumination of the working electrode, 25 the applied potential and signal measurement e) current measuring means to measure the photocurrent at the working and counter electrodes f) analysis means to derive a measure of oxygen demand from the measurements made by the photocurrent measuring means. 30 Preferably a reference electrode is also located in the measuring cell and the WO 2010/132957 PCT/AU2010/000621 8 working electrode is a nanoparticulate semiconductor electrode preferably titanium dioxide. The flow rate is adjusted to optimise the sensitivity of the measurements. This cell design is based on that disclosed in application W02004/088305 (marketed as PeCOD*) with means to store the organic/electrolyte solution. The 5 sample collection device preferably includes a filter to remove any large particulates or precipitated substances that may interfere with the operation of the cell. The method of this invention is particularly applicable to measurement of COD/organic content in industrial outflows to sea, in power plant cooling water, 10 and shipping waste water. Detailed Description of the Invention Preferred embodiments will be described with reference to the drawings in which: Figure 1 illustrates the effect of Ag 2 0-treatment on TiO 2 sensors, for analysing 15 chloride containing samples; Figures 2 and 3 further illustrate the effect of either Ag 2 0- or PdO-treatment on TiO 2 sensors, for chloride containing samples; Figure 4 illustrates a suite of highly linear calibration functions for chloride containing standard solutions generated for an Ag 2 0-treated-TiO 2 sensor; 20 Figures 5 and 6 illustrate the improved linearity obtained for chloride-free calibration standard solutions when using Ag 2 0-treated-TiO 2 as opposed to untreated TiO 2 . Methods for introducing AgOto the TiO 2 sensor matrix Standard nanoparticulate TiO 2 sensors for PeCOD* analysis were modified by 25 deposition of Ag 2 0 particles onto the TiO 2 . Two simple methods of photodeposition were employed and both gave similarly successful results for the PeCOD* analysis of chloride containing samples. In the first method of Ag 2 0 deposition, a few drops of AgNO 3 solution were added to a NaCl solution, forming a suspension of very fine AgCl particles. Before this 30 white coloured AgCl suspension could agglomerate and precipitate out, a small quantity was placed onto the thin film TiO 2 sensor surface. The surface was then irradiated for a few minutes with UV-radiation (or natural sunlight), causing WO 2010/132957 PCT/AU2010/000621 9 photoreduction of the Ag*' to fine Ago metal particles, which in turn were immediately hydrolysed to Ag 2 0 (Ag ") particles. The Ag 2 0 nanoparticles are deposited onto the TiO 2 surface and become embedded in the porous nanoparticulate TiO 2 matrix. The residual fluid was washed off the now Ag 2 0 5 treated TiO 2 sensor, which could be immediately installed into a PeCOD* instrument for use in analysis. In the second method of Ag 2 0 deposition, a few drops of AgNO 3 solution were added directly to the TiO 2 sensor surface, then the excess fluid sponged off, leaving the TiO 2 thin film matrix moist with AgNO 3 . The sensor was then placed 10 under a UV-lamp (or direct sunlight) for a few minutes, long enough to photoreduce the Ag*' to Ag 0 metal particles, which are in turn hydrolysed to Ag 2 0 nanoparticles distributed evenly through the TiO 2 sensor matrix. As above, once the residual fluid was washed off the Ag 2 0-treated sensor could be installed into a PeCOD* instrument for use in analysis. Sensors prepared using either of these 15 simple methods were used to provide the Ag 2 0/TiO 2 sensor data illustrated in Figures 1-6. It would be convenient if the Ag 2 0 could be introduced to the TiO 2 -water colloid system subsequently used to fabricate the thin layer sensor, rather than onto the already prepared, immobilised TiO 2 thin layer. However, the high temperature 20 (700 0 C) required to immobilise and calcine the TiO 2 colloid to yield the optimal anatase:rutile ratio renders this impossible. Ag 2 0 decomposes into Ago (metal) and oxygen gas at a temperature well below 700*C. The resulting thin film is a mixture of TiO 2 and Ag 0 nanoparticles. Ag metal does not provide the chloride resistance properties in the sensor that Ag 2 0 provides. 25 In Figure 1, the effect of treating a sensor with silver(l) oxide is illustrated over the chloride range 0 < [C-] < 100 mg/L. In Figures 2 and 3, the range of chloride concentrations over which the Ag 2 0/TiO 2 sensor behaviour was tested was increased to the much broader range 0 < [CI] < 4000 mg/L. Figures 2 and 3 differ only in that the x-axis is linear in Figure 2 and is logarithmic in Figure 3. With 30 reference to Figures 2 and 3, a set of reference standard solutions containing a fixed 120 mg/L COD and chloride variously at [C-] = 0, 20, 50, 100, 200, 500, 1000, 1500, 2000 and 4000 mg/L was analysed with an Ag 2 0-treated TiO 2 sensor. The chloride-free 120 mg/L COD solution was first used to calibrate the PeCOD* WO 2010/132957 PCT/AU2010/000621 10 instrument. The instrument's COD readings declined with increasing [Cl] at a rate of d[COD]/d[Cr] = -0.031 (mg/L COD)/(mg/L C-). For reference, the corresponding chloride sensitivity gradient (illustrated on Figures 2 and 3 with an unbroken line) observed for a non-Ag 2 0-treated sensor is d[COD]/d[CI-]= -1.23 (mg/L 5 COD)/(mg/L Cr)). From this we conclude that the Ag 2 0-treatment is capable of reducing the sensor's vulnerability to chloride signal suppression interference by a factor of approximately 40. Ag-treatment of TiO 2 shifts onset of chloride tailing interference to much 10 higher FCT When using an untreated TiO 2 sensor we observed that the oxidation profiles are seriously distorted by "tailing" due to chloride interference even at low chloride concentrations, [Cl] < 100 mg/L. However, when using an Ag 2 0-treated TiO 2 sensor this tailing interference effect was not seen until much higher chloride 15 levels, [C-] > 800 mg/L. We conclude that Ag 2 0-treatment of a TiO 2 sensor significantly reduces the sensor's vulnerability to the chloride tailing interference effect. Extensive subsequent testing in the laboratory has confirmed this result. Ag-treatment of TiO 2 linearises the PeCOD* response function 20 For normal PeCOD* analysis of non-chloride-containing solutions using a non Ag 2 0-treated sensor, it has been found that the instrument response function, (i.e., the Qnet vs. COD function) is significantly nonlinear. Although the degree of nonlinearity is not large, it may nevertheless result in systematic COD measurement errors of 5-10 % if a conventional straight-line, 2-point, zero and 25 span calibration protocol is employed. This systematic COD measurement error can be effectively removed by using a more sophisticated 3-5 point calibration protocol; however this introduces a significant degree of inconvenience. It would be better still to remove the nonlinearity at its source, than to work around it using sophisticated calibration protocols and polynomial arithmetic. 30 We have found, rather unexpectedly, that treating the TiO 2 sensors with silver(l) oxide results in a linear PeCOD* instrument response function. This applies to PeCOD* Ag 2 0-treated TiO 2 sensor analysis of both chloride-free and chloride containing samples, up to at least the level [Cl] = 700 mg/L. Figure 4 illustrates a WO 2010/132957 PCT/AU2010/000621 11 suite of highly linear calibration functions generated for an Ag 2 0-treated-TiO 2 sensor. Six sets of reference standard COD/C[ solutions were prepared. Set (a) consisted of solutions containing [COD] = 0, 10, 20, 50 and 75 mg/L COD and a chloride content fixed at [Cr] = 0 mg/L. Similarly, set (b) was the same except its 5 fixed chloride content was [C[] = 50 mg/L; and so on for set (c) [Cr] = 100 mg/L; set (d) [C[] = 300 mg/L; set (e) [Cl] = 400 mg/L; and set (f) [C-] = 500 mg/L. The correlation coefficient, R 2 , yielded for the straight line of best fit for each of the 5 point calibration data sets were (to the fourth decimal place): (a) 1.0000; (b) 0.9997; (c) 0.9996; (d) 0.9995; (e) 0.9996; and (f) 0.9994. 10 In Figure 5 a set of five COD (present as glucose) calibration reference standards at 0, 20, 40, 60 and 75 mg/L were analysed with a TiO 2 sensor (open circles, dotted line) and an Ag 2 0-TiO 2 sensor (filled circles, unbroken line). The normalised (at [COD] = 75 mg/L) instrument response is plotted against [COD]. The correlation coefficients for the straight line of best fit for the TiO 2 and Ag 2 0 15 TiO 2 generated data were R2 = 0.9977 and 0.9997, respectively. Figure 6 illustrates the data yielded where the COD was present as sorbitol. The correlation coefficients for the straight line of best fit for the TiO 2 and Ag 2 0-TiO 2 generated data were R2 = 0.9989 and 1.0000, respectively. Other test compoundsshowed much the same relative behaviour for TiO 2 and Ag 2 0-TiO 2 20 sensors. Ag-treatment of TiO 2 decreases sensor sensitivity to interference from dissolved CO 2 A further advantageous property acquired by the TiO 2 sensors by treating them 25 with Ag 2 0 was that they became drastically less sensitive to interference from presence of dissolved carbon dioxide in the analysed sample. Carbon dioxide

(CO

2 ), an inorganic species, is present in ambient air at a molar mixing ratio of approximately 380 pmol/mol. A water sample held in a container open to the ambient atmosphere will, over hours and days, gradually absorb CO 2 until it 30 reaches an equilibrium with the air CO 2 content. CO 2 dissolved in the water sample will in turn reach an equilibrium within the fluid between the several species CO2(aq), HCO3~(aq) and C03 2 -(aq) depending on the temperature and pH. The bicarbonate (HC03-(aq)) and carbonate (C03 2 -(aq)) ions can be readily oxidised WO 2010/132957 PCT/AU2010/000621 12 at the UV-irradiated TiO 2 surface to give a spurious (because not derived from the oxidation of an organic species) contribution to the Qnet signal collected by the PeCOD* instrument. This is generally not a significant problem if normal laboratory sample handling practices are followed and samples are stored in 5 closed containers when not in use. However, the interference from CO 2 can be a significant source of measurement uncertainty when analysing relatively clean water containing low levels of COD (< 20 mg/L). It has been found that Ag 2 0 treated TiO 2 sensors are quite insensitive to the presence of dissolved CO 2 in water samples, virtually eliminating this as a source of measurement uncertainty. 10 Ag-treatment of TiO 2 decreases the analysis time using pulsed-UV irradiation for highly saline samples It has been demonstrated already that treatment of TiO 2 sensors with Ag 2 0 extends the range of COD analysis in chloride-containing matrices when using 15 continuous UV-irradiation from 0 mg/L < [C-] < 80 mg/L to at least 0 mg/L < [Cl] < 700 mg/L. At chloride concentrations beyond 700 mg/L it may still necessary to use pulsed-UV irradiation to circumvent chloride interference. This allows analysis for COD in water and wastewater samples containing chloride at concentrations of at least 21,000 mg/L (the level typically found in seawater). We have found that 20 the Ag 2 0-treatment of TiO 2 sensors also brings an improvement to this type of analysis. Pulsed-UV analysis proceeds more rapidly, by 20-30 %, when using an Ag 2 0-TiO 2 sensor as opposed to a TiO 2 sensor. This yields a significant time saving for users of this method. 25 Possible mechanism of the advantageous effects arising from Ag 2 0 treatment of the TiO 2 sensor The mitigation of interference effects arising from dissolved C[ and CO 2 on a TiO 2 sensor, by treating it with silver(l) oxide, have not been reported in the scientific literature before and are not well understood. The best we can do at present is 30 speculate on what the mechanisms of these effects may be. It is believed that the photocatalytic oxidative action of TiO 2 occurs predominantly at the solid-liquid interface, i.e., at the TiO 2 surface, through oxidative attack by the photohole (hvab) in the TiO 2 valence band. Species having a propensity to adsorb onto the TiO 2 WO 2010/132957 PCT/AU2010/000621 13 surface are typically oxidised more efficiently than those species that do not adsorb. We suspect that Cr, and dissolved C02 are species that do adsorb to the TiO 2 surface. In doing so they may compete with the organic species in the solution for the photoactive TiO 2 sites where oxidation may occur. This may result 5 in interference with the oxidation of the organics, such as we have observed using untreated TiO 2 sensors. We suspect that deposition of Ag 2 0 particles onto the TiO 2 changes the nature of the TiO 2 surface in such a way that C1, HCO3 and C032- are less prone to- adsorb on this surface, and thereby less able to interfere in the COD measurement using 10 this approach. We suspect that the deposition of Ag 2 0 onto the TiO 2 changes the surface charge of the combined Ag 2 0-TiO 2 surface, making it more negatively charged. Plausibly, the anions in solution perceive the Ag 2 0-TiO 2 surface as being more electron rich than they would a bare TiO 2 surface and are less likely to adsorb to the former because of relatively repulsive electrostatic forces. If Cr, 15 HC03- and C032- anions are significantly less likely to adsorb to a Ag 2 0-treated TiO 2 surface than to a TiO 2 surface, then we can expect the Ag 2 0-TiO 2 sensor to be subject to much less interference from these anions in the process of COD analysis. This is also a plausible explanation for the Ag 2 0-treatment of a TiO 2 sensor 20 linearising the previously nonlinear response function, if we consider that this nonlinearity may be due, indirectly, to HC03- and C032-. Recall that the mineralisation (i.e., exhaustive oxidation) of the organics results in C02 as a major product. This product C02 will to some extent dissolve in the sample and move towards establishing a C02(aq)/HCO3~(aq)/CO3 2 -(aq) equilibrium. This product HC0 3 ~ 25 and C032- may be oxidised again at the TiO 2 surface to give C02, and a spurious contribution to the measured COD. Thus, the higher the concentration of organics initially in the cell, the more likely there is to be established this positive feedback loop based on the C02, HC03- and C032- oxidation-reduction cycle in solution. This will manifest as the sort of concave upward nonlinear response function we 30 observe for bare TiO 2 sensors, as illustrated in Figures 5 and 6. To the extent that Ag 2 0 treatment mitigates against the adsorption of HCO3- and C032- anions on the Ag 2 0-TiO 2 surface, the nonlinearity should diminish.

WO 2010/132957 PCT/AU2010/000621 14 TiO 2 sensor treatment using oxides of metals other than silver Silver is an element belonging to the Group 11 metals of the Periodic Table of Elements. It also belongs to a set of metal elements, the noble metals, from the same part of the Periodic Table, that frequently exhibit similar chemical properties. 5 This group includes another Group 11 element Gold (Au), and the two Group 10 elements Platinum (Pt) and Palladium (Pd). The scientific literature reflects that treatment with these three other metals frequently lends similar properties to nanoparticulate TiO 2 (and other semiconductors) as does treatment with silver. It might reasonably be expected that, analogously, treatment of TiO 2 sensors with 10 the oxides of Au, Pt or Pd may lend similar advantageous properties to TiO 2 sensors as does treatment with Ag 2 0. With this in mind palladium(II) oxide was deposited onto TiO 2 sensors in order to test these sensors for enhanced chloride resistance. With reference to Figures 3 and 4, the chloride resistance of the PdO TiO 2 sensor was characterised in the same way as it was for the Ag 2 0-TiO 2 and 15 untreated TiO 2 sensors, as described above. While the profile of the measured COD vs. [CI] function for the PdO-TiO 2 sensor was quite different from that of Ag 2 0-TiO 2 sensors, it did show very significantly enhanced chloride interference resistance with respect to untreated TiO 2 . Indeed, the PdO-TiO 2 sensor continued to function as a COD sensor at 4000 mg/L chloride, with a much smaller degree of 20 signal suppression than that shown by Ag 2 0-TiO 2 sensors. From this it follows that PdO-TiO 2 sensors will facilitate the analysis of COD in seawater (containing -21,000 mg/L C-), requiring only a relatively modest 5x dilution of the sample prior to PeCOD analysis. Preferably, for such analysis, the instrument would be calibrated with appropriately saline standard solutions rather than chloride-free 25 solutions. Several methods of Pd and PdO deposition are well described in the literature. These include photodeposition approaches with starting materials such as PdNO 3 or other soluble palladium salts. There are also thermal deposition approaches, using similar starting materials. In the present case, an even simpler deposition method was used. Commercially sourced PdO nanoparticles were 30 added to the 5% TiO 2 /water colloid system normally used for thin film TiO 2 sensor fabrication, and well mixed. The mass ratio PdO:TiO 2 in the resulting mixed nanoparticle suspension was approximately 15%. This mixture was used to fabricate slides by the usual dip-coating method followed by calcination at 700 0

C,

WO 2010/132957 PCT/AU2010/000621 15 as disclosed in W02009/062248. Ag-treatment may mitigate against interference from halide ions other than chloride Chloride is an ionic species belonging to the halide group (from Group 17 of the 5 Periodic Table of Elements). The other halide ions Fluoride (F-), Bromide (Br) and Iodide (l-) frequently exhibit very similar chemical properties to those of Chloride. We claim that from this it follows that Ag 2 0-treatment of TiO 2 sensors may mitigate against interference by these other halide ions in the same way it mitigates against interference by chloride. 10 AgO-treatment of TiO 2 sensors used in amperometric measurement of COD While the present application has been described with reference to a previously disclosed coulometric method for measuring COD (WO 2004/088305), the same application also discloses an amperometric COD measurement method. 15 Furthermore, other related applications disclose amperometric methods for COD measurement (W02008/077192, W02008/077191, and AU 2010900885). These methods are also based on the principle of photocatalytic oxidation of organics on a nanoparticulate TiO 2 surface. The advantages of treating the TiO 2 sensor with Ag 2 0 would also apply to these methods. 20 Those skilled in the art will realise that the present invention provides a robust analytical tool that can provide accurate measurement of COD in a short time without interference from competing species such as chloride. It will be evident to those practiced in the art that this novel new method also has 25 potential application for other interfering species, such as other soluble counter ions (including bromides, iodides, sulphates, phosphates etc.) Those skilled in the art will also realise that this invention may be implemented in embodiments other than those described without departing from the core teachings of the invention. 30 PeCOD* is a trademark of Aqua Diagnostic Holdings PTY LTD

Claims (8)

1. A method of determining chemical oxygen demand in water samples containing fluoride, bromide or iodide and/ or chloride ions by a 5 photoelectrochemical method in which the photo electrode is a titanium dioxide sensor which has been treated with a noble metal compound.

2. A method of determining chemical oxygen demand in water samples containing fluoride, bromide or iodide and/ or chloride ions by a 10 photoelectrochemical method in which the photo electrode is a titanium dioxide sensor in which a noble metal oxide is incorporated into the nanoparticulate TiO 2 material matrix by means of photodeposition, thermal deposition or chemical synthesis. 15

3. A method as claimed in claim 1 or 2 in which the noble metal is selected from silver or palladium.

4. Water quality assay apparatus for determining oxygen demand of a water sample which consists of 20 a) a flow through measuring cell b) a photoactive titanium dioxide working electrode incorporating a noble metal oxide and a counter electrode disposed in said cell, c) a UV light source, adapted to illuminate the photoactive working electrode either continuously or in pulses 25 d) control means to control the illumination of the working electrode, the applied potential and signal measurement e) current measuring means to measure the photocurrent at the working and counter electrodes f) analysis means to derive a measure of oxygen demand from the 30 measurements made by the photocurrent measuring means. WO 2010/132957 PCT/AU2010/000621 17

5. Water quality assay apparatus as claimed in claim 4 in which a noble metal oxide is incorporated into the nanoparticulate TiO 2 material matrix by means of photodeposition, thermal deposition or chemical synthesis. 5

6. Water quality assay apparatus as claimed in claim 5 in which the noble metal oxide is selected from silver or palladium.

7. Water quality assay apparatus as claimed in claim 6 in which the noble metal oxide is silver(l) oxide and the Ag 2 0:TiO 2 density ratio of the 10 resulting Ag 2 0-TiO 2 composite material is controlled by manipulating the deposition parameters.

8. Water quality assay apparatus as claimed in claim 6 in which the noble metal oxide is palladium(II) oxide and the PdO:TiO 2 density ratio of the 15 resulting PdO-TiO 2 composite material is controlled by manipulating the deposition parameters. 20